Package 'phenorice'

phenorice

Description

phenorice model to detect rice from satellite reflectance data

References

Boschetti, M., L. Busettoa, G. Manfrona, A. Laborte, S. Asilo, S. Pazhanivelan & A Nelson, 2017. PhenoRice: A method for automatic extraction of spatio-temporal information on rice crops using satellite data time series. Remote Sensing of Environment 194: 347-365

---

Smooth and/or fill time series data

Description

Fill missing values (pheno_fill_VI) or fill and smooth (pheno_filter_VI) a vector with time series data

Usage

pheno_fill_VI(x, method = "natural")
pheno_filter_VI(x)

Arguments

x	numeric vector
method	character. Passed to spline

Value

numeric

---

Default parameters

Description

Get default parameters for the phenorice model

Usage

getPars()

Details

	Parameters	Explanation
	evi_meanth	threshold for annual mean EVI
	evi_maxth	threshold for maximum evi
	evi_minth	threshold for minimum evi
	pos_start	start of heading
	pos_end	end of heading
	vl1	the shortest vegetative growth length
	vl2	the longest vegetative growth length
	winfl	half period of flooding
	minndfi	threshold for ndfi
	windecr	period after maximum EVI
	decr	percent decrease of EVI after EVI maximum
	tl1	the shortest total growing length
	tl2	the longest total growing length
	lst_th	the minmum temperature for planting

Value

list

Examples

p <- getPars()
str(p)

---

Date of maximum vegetation index

Description

Find the time-step of a prominent local maximum in a vegetation index time series within a given period.

Usage

maxDate(vi, vi_maxth = 0.3, pos_start = 15, pos_end = 40)

Arguments

vi	numeric vector with vegetation index values
vi_maxth	numeric. Threshold for a valid local maximum after subtracting the series minimum
pos_start	integer. First time-step (index) of the search window
pos_end	integer. Last time-step (index) of the search window

Details

The background (minimum) value is subtracted from vi. Local maxima are identified, filtered by vi_maxth and by the window [pos_start, pos_end], and checked for an increase before and decrease after the peak. If more than one candidate remains, the date closest to pos_end is returned.

Value

integer index of the selected maximum, or integer(0) if none is found

Examples

evi <- readRDS(system.file("evi.rds", package="phenorice"))
evi <- pheno_fill_VI(evi)
maxDate(evi)

---

R implementation of the phenorice model

Description

Run the phenorice model

Usage

phenorice(evi, ndfi, p, lst)

Arguments

evi	numeric, rasterstack or SpatRaster with evi values
ndfi	numeric, rasterstack or SpatRaster with ndfi values
p	list with named parameters based on phenorice method
lst	numeric, rasterstack or SpatRaster with lst values

Details

	Parameters	Explanation
	evi_meanth	threshold for annual mean EVI
	evi_maxth	threshold for maximum EVI
	evi_minth	threshold for minimum EVI
	pos_start	start of heading
	pos_end	end of heading
	vl1	shortest vegetative growth length
	vl2	longest vegetative growth length
	winfl	period for flooding
	minndfi	threshold for NDFI
	windecr	period after EVI maximum
	decr	percent decrease of EVI after EVI maximum
	tl1	shortest total growing length
	tl2	longest total growing length
	lst_th	minmum land surface temperature for planting, needs checkLST=TRUE

Value

vector

References

Boschetti, M., L. Busettoa, G. Manfrona, A. Laborte, S. Asilo, S. Pazhanivelan & A Nelson, 2017. PhenoRice: A method for automatic extraction of spatio-temporal information on rice crops using satellite data time series. Remote Sensing of Environment 194: 347-365;

Busetto, L., Zwart, S.J. and Boschetti, M., 2019. Analysing spatial?temporal changes in rice cultivation practices in the Senegal River Valley using MODIS time-series and the PhenoRice algorithm. International Journal of Applied Earth Observation and Geoinformation 75: 15-28.

Wang, H., A. Ghosh, B.A. Linquist, R.J. Hijmans, 2020. Satellite-based observations reveal effects of weather variation on rice phenology. Remote Sensing 12(9):1522

Examples

# Get index for one pixel
evi <- readRDS(system.file("evi.rds", package="phenorice"))
ndfi <- readRDS(system.file("ndfi.rds", package="phenorice"))
lst <- readRDS(system.file("lst.rds", package="phenorice"))

# Smooth each index
evi <- pheno_filter_VI(evi)
ndfi <- pheno_fill_VI(ndfi)
lst <- pheno_fill_VI(lst)

# Get parameters
p <- getPars()

# Run and get result (planting date, date of max EVI value, flowering date, heading date, harvest date) for tha pixel 
re <- phenorice(evi, ndfi, p, lst)

# Change parameters
p['evi_minth'] = .2
re <- phenorice(evi, ndfi, p, lst)


f <- system.file("ricevi.rds", package="phenorice")
d <- readRDS(f)
evi <- unlist(d[1,2:47])
ndfi <- unlist(d[1,48:93])
lst <- unlist(d[1,94:139])
p <- getPars()

phenorice(evi, ndfi, p, lst)


x <- matrix(NA, 25, 5)
for (i in 1:25) {
	evi <- unlist(d[i,2:47])
	ndfi <- unlist(d[i,48:93])
	lst <- unlist(d[i,94:139])
	x[i,] <- phenorice(evi, ndfi, p, lst)
}
x

---